Lanthanide interactions with nitrotyrosine. Specific ... - ACS Publications

May 7, 2018 - Lanthanide Interactions with Nitrotyrosine. A Specific. Binding Site for Nuclear Magnetic Resonance Shift. Probes in Proteins. Timothy D...
0 downloads 0 Views 1MB Size
6562 (14) K. Wuthrich, Struct. Bonding(Berlin), 8, 53 (1970). (15) (a) G. Redfield and R. J. Gupta, CoM Spring Harbor Symp. Ouant. Bo/., 36, 405 (1971): (b) C. C. McDonald and W. D. Phillips, Biochemistv, 12, 3170 (1973). (16) (a) R. M. Keiier, G. W. Pettigrew. and K. Wuthrich, FEBS Left., 38, 151 (1973): (b) C. C. McDonald, W. 0. Phillips, and J. LeGali, Siochemisfry, 13, 1952 (1974): (c) G. M. Smith and M. D. Kamen, Roc. Natl. Acad. Sci. U.S.A.,71, 4303 (1974). (17) A. Kowaisky, Biochemistry, 4, 2382 (1965). (18) H. Davenport, Nature(London), 169, 75 (1952). (19) F. Sherman, J. W. Stewart, J. H. Parker, E. Inhaber, N. A. Shipman, G. J. Putterman, R. L. Gardisky, and E. Margoliash. J. Biol. Chem., 243, 5446 (1968). (20) H. Muir and A. Neuberger. Biochem. J., 45, 163 (1949). (21) H. Budzikiewicz, C. Djerassi. and D. Williams, "Mass Sepctrometry of Organic Compounds", Holden-Day, San Francisco, Calif.. 1967, pp 628-629. (22) A. N. H. Yeo, Chem. Commun., 1154 (1970). (23) T. Popper and H. Tuppy, Acta Chem. Sand., 17, S47 (1963). (24) (a) H. Theorell. Biochem. Z.,301, 201 (1939); (b) D. Theodoropouios and I. Souchleris, J. Org. Chem., 31, 4009 (1966); (c) H. Gnichtel and W. Lautsch.. Chem. Ber., 98, 1647 (1985). (25) (a) H. Theorell. Biochem. Z., 295, 207 (1936): (b) ibid.. 298, 242 (1938); (c) ibM., 301, 201 (1939); (d) Enzymologia, 4, 192 (1937). (26) J. B. Paine and D. Dolphin, J. Am. Chem. SOC., 93,4080 (1971). (27) R. Bonnett, I. A. D. Gale, and G. F. Stephenson, J. Chem. SOC.C,1600 (1966). (28) S.J. Jacobson, C. G. Willson, and H. Rapoport, J. Org. Chem., 39, 893 (1974). (29) G. W. Kenner. S. W. McCombie, and K. M. Smith, Justus Liebigs Ann. Chem., 1362 (1973). (30) R . P.Carr. A. H. Jackson, G. W. Kenner. and G. S.Sach. J. Chem. SOC.

C.487 (1971). (31) The small but distinct differences between synthetic and natural porphyrin c will be discussed in a future publication. (32) Solvent evaporations were carried out in vacuo using a Berkeley rotary evaporator, ultraviolet and visible spectra were measured on a Cary Model 14 spectrometer, NMR spectra were taken on Varian Model T-60. HA-100, and HR-220 spectrometers using internal Me& and mass spectra were obtained on CEC 103 and llOB and A.E.I. MS9 spectrometers. The 200-tube automatic Craig type distribution train was manufactured by H. 0. Post Scientific instrument Co. Radioactivity determinations were made with Packard TriCarb equipment; sample oxidation in sample oxidizer, Model 305, and counting in liquid scintillation counter, Model 3375. Elemental analyses were performed by the Analytical Laboratory, Department of Chemistry, University of California, Berkeley. (33) V. du Vigneaud, L. F. Audreith, and H. S. Loring, J. Am. Chem. SOC., 52, 4500 (1930). (34) D. Seebach, B. Erickson, and G. Singh, J. Org. Chem.. 31, 4303 (1966); A. Vallet, A. Janin, and R. Romanet, J. Labelled Compd., 4, 299 (1968). (35) C. S.Marvel and F. Hager, "Organic Syntheses", Collect. Voi. I, Wiley. New York. N.Y., 1944, 248. (36) The preparation of [2- H,]ethanoi has been reported via LiAID4 reduction of ethylene oxide by E. Bengsch and M. Corval, Bull. SOC.Chim. Fr., 1867 (1983). In our hands this procedure resulted in significant quantities of both dideuterio and unlabeled products. The isotopic scramble is difficult to detect by NMR or by mass spectral analysis of the alcohol because pressuredependent intermolecular proton transfer in the mass spectrometer and the natural isotopic mixture of oxygen renders the M+ region complex. In the derived iodlde, these problems are both obviated and the molecular ion region of the iodide mass spectrum is uncomplicated and accurately reveals this isotopic mixture. (37) M. M. Kreevoy and M. A. Turner, J. Org. Chem., 30, 373 (1965).

B

Lanthanide Interactions with Nitrotyrosine. A Specific Binding Site for Nuclear Magnetic Resonance Shift Probes in Proteins Timothy D. Marinetti, Grayson H. Snyder, and Brian D. Sykes* Contributionfrom the Department of Chemistry, Harvard University, Cambridge, Massachusetts 02138. Received March 20, 1975

Abstract: A 100-MHz ' H N M R study of the relaxation and chemical shifts induced by the interaction of Eu(III), Pr(III), Gd(lIl), and La(II1) with N-acetyl-L-3-nitrotyrosine ethyl ester has been undertaken to characterize the nitrotyrosine residue as a potential specific lanthanide metal binding site in proteins. The N M R data, together with independent measurements of the metal-ligand stability constants, give the parameters of the dipolar interaction. These indicate a significant contribution from nonaxially symmetric terms. In addition, a nonnegligible contact shift is present for Eu( 111).

The utility of the lanthanide shift probes as an aid in determining molecular structure and conformation in solution has been well documented over the last several years.',2 Most of the work to date has involved the use of "shift reagents'' [e.g., Ln(fod)3] in organic solvents. Increasingly more interest is being shown in using the aquo cations themselves, or their EDTA complexes as aqueous shift reagents, for structure determination in aqueous solution. In particular, several molecules of biochemical interest have been studied including mononucleotide^,^^^ dinucleotide^,^ and Recently the Oxford enzyme group has been investigating the use of the lanthanide binding to a specific site on the protein lysozyme to determine the three-dimensional structure of the protein in solution with minimal reference to the known X-ray A limitation of the extension of this technique to other proteins is that the protein must possess a single specific naturally occurring binding site. Although the lanthanides are reported to bind to several proteins such as t h e r m ~ l y s i n ,a-amylases," ~~ trypsinogen,I2 * Addres\ correspondence to this duthor at Department of BiochemiWy. The L n l v e r s i t y of Aibertd Edmonton. Aibertd. Cdndda T6G-ZH7

Journal of the American Chemical Society

and ferredoxin,I3 this will obviously not be true for all proteins of interest. One way to circumvent this difficulty is the incorporation of a binding site into the p r ~ t e i n W . ~e have therefore investigated the chelating group nitrotyrosine which can be introduced with relative ease into a large number of proteins.I4 In many of these proteins, nitration occurs a t only one tyrosine, hence generating a single binding site for the metal ion. This paper reports the study of the interaction of Pr(III), Eu(III), and Gd(II1) with the model peptide N-acetyl-~-3nitrotyrosine ethyl ester by measurement of the induced N M R shifts, line widths, and spin-lattice relaxation times. I n addition, by potentiometric titration, the binding constants of these metals and La(III), a diamagnetic analog of the 4f series, have been determined. With known binding constants, the equilibrium composition of the samples used in N M R can be calculated, and, hence, data involving a change in a spectral parameter as a function of the concentration of paramagnetic metal ion can be analyzed to extract the parameters for the metal-ligand complex. Using the isotropic relaxation probe Gd(II1) and the known X-ray structure of nitrophenol, the location of the metal binding

/ 97:22 / October 29. I975

6563

site and the atomic coordinates in the metal complex were calculated. Since the nitrotyrosine ring is rigid, the ring protons provide a test of the validity of the dipolar model in predicting lanthanide-induced shifts. Our analysis of the N M R shifts indicates that the dipolar model works very well for Pr(II1) with the contact contribution estimated to be less than 2% of the smallest shift. For Eu(III), the results indicate that a significant contact interaction is involved, but the dipolar contribution is still dominant. Operational parameters have been obtained for use of the nitrotyrosine residue as a chelating site in proteins. Using these parameters, the observed lanthanide-induced shifts in the N M R spectrum of a nitrated protein can give information about the placement of the observed residue relative to the metal. Finally, a general theoretical discussion shows that the principal axes of the dipolar interaction cannot be unambiguously determined from solution N M R measurements. Further, axial symmetry will be indicated by internal shift ratios only if the ratios are constant to within a few percent. Deviations greater than this imply that nonaxial terms are contributing to the shifts. Theory. Following Bleaney15 and Horrocks,I6 we treat the lanthanide dipolar shifts as arising from anisotropy in the magnetic susceptibility tensor, X, which is the population average of the magnetic interactions of the 4f electrons in the thermally occupied J sublevel of the ground term. When electron spin relaxation is fast compared to nuclear relaxation, the electronic magnetic moment seen by the nucleus is averaged over all thermally accessible electronic states. We consider a ligand molecule in some molecular frame, which can be related to the lab frame by a coordinate rotation R. Assuming the lanthanide-induced shift is small, the nuclear spins will be quantized along the main field H. In the molecule fixed frame the magnetic moment of the nucleus becomes R-l.1, and the field is similarly R-‘*H. The magnetic moment of the metal ion in the molecular frame is pe = X*R-’*H.The interaction energy is the magnetic interaction which, for a nuclear dipole at r in the molecular frame, can be written as

X D w1.p2r-~- 3(p,.r)(p2.r)r-5

(1)

with p~ = yhR-’-1, p2 = pe = X-R-I-H. This can be rewritten in tensor notations: X D = p 1 , D - p ~where D is the dipole-dipole interaction tensor.]’ Substituting for p~ and p~

X D= ( y h R - l . ~ ) . ~ ~ -=~ - l - ~ yhl.RaDX*R-l*H E 7fiI.T.H

(2)

where T = R0DXR-I. This is formally analogous to a chemical shielding tensor, as can be seen by writing out the complete Hamiltonian for the nuclear spin I

X = -yhI-(l

- u)-H + y h I - F H

(3) where u = R*d*R-I and u’ is the screening tensor in the molecule fixed frame. Averaging over all possible molecular orientations replaces the tensor quantities by their scalar average (I/3)tr(q and hence the effect of the metal ion is to shift the nuclear resonance by an amount (-!/3)tr{q. By analogy with the screening tensor this is dimensionless and represents the relative shift AW/OO.Since the trace is invariant under similarity transform, such as spatial rotation, we are free to choose our coordinate system. Taking it as the principal axis system (in which Xis diagonal)

From the explicit forms of the dipole tensor, and using the fact that t r l q = 0, the above expression reduces to

+

(Ao/wo) = I / ~ F ~ [ ( X , , % ) ( 3 cos2 8’ - I ) (Xxx - X,,) ( sin2 8‘ cos Q ’ ) ]

(5)

with = (I/3)tr{X),and the polar coordinates of the nucleus (r,O’,@’) are in the principal axis system. The above is identical with Bleaney’s result,I5 except for the factor of N , which is absorbed into our X tensor. Using normalized spherical harmonicsL8 1hr-3(3cos2 8’

- 1)

= r-3c20(8’,4’)

1/2r-3(sin2 0’ COS 24’) = q r - 3 ( ~ 2 2 ( 0 ’ , @ ’+) C2-2(8’,4’))

(6)

If we start in some arbitrary coordinate system, in which we know the coordinates (r,O,@)of all atoms of interest and wish to know what the shift is, the procedure involves calculating the geometrical functions and seeing how well the data are fit to a simple linear combination of them. Let the real principal axis be at Euler angles R A, p , u from the initial coordinate system. To calculate C20(8’,4’), C22(8’,@’),and C2-2(8’,4’) in the principal axis, we calculate values (in the starting system) of the C2,(8,4) and note that the transformation properties of the Ck,(O,@) under rotation are well characterizedI8

Ckn(@’,@’) = C Dmnk( fl )Ckm(8.4) m

(8)

where (e,@) are the coordinates of the point in the arbitrary coordinate system and Dmnk(f2)is a Wigner rotation matrix element. Using this, and the fact that r is invariant under rotation, the shift becomes (Aw/wo) =

m

+

[ ( X , , - X)Dmo2(R)

q w x x- X,,)(Dm22(f2) + Dm-22(Q))lC2m(8,@) (9)

Not knowing the principal axis transform angles A, p, u, we select arbitrary angles C Y , b, y andfit the data to a function of the form ( A o / w o )= r 3 [ A I D ~ O ’ ( ~ . + P ,A~2 )( D m 2 2 ( a , P , r + )

tm:

D m - 2 2 ( d 3 . r ) 1I C Z m ( 8 . 4 )

( 10)

the success of which will be determined by how well the dipolar interaction can account for the observed shifts. Since the parameters A I , A2, C Y , P, and y must be the same regardless of r , 0, @, Le., for any nucleus in the paramagnetic complex, then each of the coefficients must separately be equal A lDmoZ(a$,r>+ A2(Dm22(CY9P,r)-3

= ( X z z - X)D~O‘(A.N~~) + - X , y ) ( D m 2 * ( ~ A V )+ Dm-22(A,KcL,.)) (1 1)

Dm-22(..b,y)) q

m

x

x

for m = 0, f l , f 2 . We have, in effect, six unknowns (A, y, v and the principal values of the susceptibility) and a system of five equations relating them to quantities A I , A2, cr, @, and y which are determined by a fit to the observed chemical shifts. The situation is therefore underdetermined and there are an infinite number of solutions satisfying eq 11.

Marinetti, Snyder, Sykes

/ Lanthanide Interactions with Nitrotyrosine

6564 One of these solutions will have A2 = 0, which represents a “pseudo axial” solution to the N M R shifts. Therefore, fitting the N M R data to an axial model is not significant in itself, since it will be possible to fit any dipolar shift with an axial expression by choice of A I ,a , /3, and y . We must ask if the resulting axes are physically reasonable or, to turn the question around, if a set of physically reasonable axes can produce values of A I and A2 where A2